Research On The Combined Forecast Analysis Model Of The Yield Of Peanut

Prediction of grain yield is one of the most important basic work of regional development planning.With the increase of population and the development of economy,the task of ensuring food security will be more serious.It is of great significance for the macro decision-making of the national economy to carry out scientific and accurate prediction of grain yield.Peanut is one of the most important crops in China,a ccurate prediction of the peanut yield can provide the basis for the formulation of relevant policies.With the application of prediction technology is more and more widely,people's demand for prediction accuracy is also increasing.At present,there are many methods to predict the grain yield,such as support vector machine(SVM),grey theory and neural network.However,the single support vector machine model and grey system model can not predict the grain yield very well,and it is more suitable to describe the general trend of grain yield.In view of the science and effectiveness of the practical application of the combined forecasting method,this paper attempts to predict its application in peanut production in Henan province Neihuang County,in order to improve the prediction accuracy of peanut production.In this paper,two methods are selected to predict the single model: exponential smoothing prediction model and GM(1,1)prediction model.Exponential smoothing method of Time series prediction method is ideal for easy operation;and the exponential smoothing method can track the change of the data,adjust the estimation of the trend of the sequence,In the short-term prediction,because the real time information is few,it is ideal to choose the exponential smoothing method;GM(1,1)prediction model is widely used in agriculture,industry and other fields because of its advantages such as less sample data,convenient operation,high accuracy and can be tested.GM(1,1)prediction model is mainly applicable to a single exponential growth data sequence.This two methods considered time series,single factor,multi factor causality,linear and nonlinear relation,etc.Therefore,this two models can be used to analyze and predict the model,and the advantages of each model can be realized,to avoid the drawbacks of a single model in terms of the relationship between factors and functions,it is helpful to improve the accuracy of prediction and the comprehensiveness of response information.